The Great Powers are in an Artificial Intelligence (AI) arms race. China publicly set a goal of becoming “the world’s primary AI innovation center” by 2030. Vladimir Putin said of AI that whoever “becomes the leader in this sphere will be the ruler of the world.”
Fortunately, the National Geospatial-Intelligence Agency (NGA) was founded on the belief that innovation is the key to our success – today and tomorrow. It’s in our DNA. Our predecessor agencies bequeathed a legacy of focusing on the customer and solving their problems. From the Cuban Missile Crisis, to mapping the Moon for NASA landings, to the hunt for Osama bin Laden, our women and men have expertly practiced our craft, embraced new technologies, pushed the frontiers of innovation, and found ways to extract every ounce of value out of every bit of data we’ve collected.
More so than any time in human history, technology is revolutionizing how we live – and how we work. Cars park themselves… and are on the verge of driving themselves. Those same cars, along with drones, planes, small satellites – and even light posts – are sensing, collecting, processing, and sharing data at a prolific rate.
The potential of that data has particular implications for the Intelligence Community (IC) – and my own agency, NGA. At this year’s GEOINT Symposium in April, I laid out how NGA must be a leader in this evolving environment and enable the IC for future success. I discussed both the art and the science of GEOINT and the critical role the human analyst must play as we adopt tools and apply automation to turn that data potential into reality.
As I see it, the GEOINT community finds itself at the same juncture the SIGINT community was at the advent of the computer age – a move from data scarcity to data abundance, and from intermittent sources to a torrent of source material collected, processed and stored by these new machines.
Some see automation as a choice that diminishes the importance of human expertise. I see that as a false choice. In fact, I see the reality of exponential automation as elevating and scaling our finite human expertise. It is not human versus machine – it is human plus machine.
To realize that independent benefit, we must continue to leverage automation and machine learning to support the continued research and development of artificial intelligence. And we must do it now.
As we continue this journey, we know we need to rely more on industry and academia than ever before – to mature our current efforts into long-lasting and accessible capabilities, but also explore and develop new capabilities together. We are making good progress. In the last year, NGA launched dozens of automation, augmentation and artificial Intelligence projects.
Recently, I announced the standup of the GEOWorks initiative. This is a new, online data sharing platform. This allows industry and academia to access unclassified data and tools and assist in building solutions we need to promote greater collaboration. This unclassified data can be used to build algorithms, participate in hackathons and compete in challenges – all with the goal of solving hard problem sets and identifying transferable tools or techniques.
We’ve also created a team of specialized professionals – our Data Corps – to help our geospatial analysts build upon their imagery foundation to include a data-centric focus. This team teaches our analysts how to pair their deep analytical expertise with the latest technologies – such as machine learning and computer vision – to address the hardest challenges facing the Intelligence Community. With these skills, these experts – our analysts – are creating new understanding and deriving more meaningful insights from our vast data sets.
And just last month we signed a cooperative research and development agreement (CRADA) with Planet, a commercial imagery provider. With this CRADA, we will explore – and improve – the speed at which the agency extracts vital information and develop analytics from imagery. This is new territory and we are really encouraged by innovative companies like Planet. And we are excited about how we can make the different pieces of this puzzle fit.
We are also working to equip our analysts with the right AI tools to help them do their jobs better and faster. Our Automation, AI and Augmentation (AAA) efforts are dedicated to meeting the needs and requirements of the GEOINT analytic community. Our first priority is to bring coherence to NGA’s existing automated tools and products, making sense of current resources, and defining future needs.
This is not the adoption of new tools and techniques simply because they are new, or shiny, or trendy. This is a focused effort, led by members of our analytic cadre, to give analysts the one thing they always want more of… time.
This is all about empowering our analysts. It’s about developing technology to best meet the needs of the geospatial intelligence officers of the 21st century. We must give them the tools to compete – and win – in today’s escalating AI arms race.
Our customers deserve – and the world demands – no less.